AI Agent Hub
Back to plugins
🖥️

dsh-pet-panel

Client Updated 2026.09.05

Run the following command in DeepSeek Harness:

dsh plugin install zw11591-sketch/dsh-pet-panel

Paste the following prompt into your AI chat to install this plugin:

The source code is available at https://github.com/zw11591-sketch/dsh-pet-panel; install it in DeepSeek Harness by running dsh plugin install zw11591-sketch/dsh-pet-panel.

About this plugin

dsh-pet-panel turns the DeepSeek Harness Web UI from a plain chat window into a workbench with personality, visibility, and collaboration. The most eye-catching entry point is a floating desktop pet that hovers above the conversation area — draggable, skin-switchable across five SVG species, and reactive to session lifecycle (busy during runs, waiting on pending, celebrating on completion). Toggle it with /pet on or /pet off, and it gives long screen sessions a lightweight visual anchor.

Beyond the pet, the plugin fills a real gap in self-service management. On the browser side you get a Skill Forge (CRUD for SKILL.md files plus natural-language skill generation via the default model), MCP server management, A2A agent registration with live online/latency status, and a team collaboration panel with @mention routing and multi-turn memory — no more digging through config files or the terminal. On the host side, the plugin exposes an inbound A2A endpoint and outbound calling tools so the model itself can discover and invoke external agents, sharing the same model, skills, and MCP context as the Web UI while keeping strict per-profile data isolation.

If you run multiple AI profiles, want to manage skills and external agents from a single browser surface, or want to spin up multi-agent group chats mid-conversation, dsh-pet-panel closes the loop from observe to manage to collaborate, removing the friction of repeatedly opening config files and terminals.

Use Cases

  • Create, edit, and delete SKILL.md files and MCP servers entirely from the browser without switching to the terminal
  • Register external A2A agents and build multi-agent group chats with @mention routing and per-agent context memory
  • Track session state at a glance via a floating desktop pet that reacts to runs, pending tasks, and completions

Best For

  • Developers running multiple AI profiles who need per-profile isolation of skills, agents, and tools
  • Team leads who want a unified browser surface to manage skills, MCP, A2A agents, and teams
  • Independent developers who work long sessions in the Web UI and want a visual companion to track conversation progress